# PKU-MARL/DexterousHands

This is a library that provides dual dexterous hand manipulation tasks through Isaac Gym

Repository: https://github.com/PKU-MARL/DexterousHands
Canonical: https://ross.abutalabs.com/products/dexteroushands
Homepage: https://pku-marl.github.io/DexterousHands/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: deep-reinforcement-learning, dexterous-robotic-hand, reinforcement-learning
Last push: 2025-02-18T13:47:57+00:00

## Health v2 (maintenance only)
Score: 35/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 7, release rhythm 35, longevity 100
- inputs: {"age_days": 1618, "days_push": 561, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1085, forks 134 (observed 2026-08-28T04:03:31.452058+00:00)

## What it is
Bi-DexHands is a Python library providing bimanual dexterous hand manipulation environments built on NVIDIA Isaac Gym for reinforcement learning research. It includes a comprehensive benchmark of RL, MARL, multi-task RL, meta-RL, and offline RL algorithms across tasks like handover, throw, and lift with thousands of YCB/SAPIEN objects.

## Use cases
- train RL agents on bimanual dexterous hand manipulation
- benchmark multi-agent RL algorithms on robotic hand tasks
- run thousands of parallel physics simulations on a single GPU
- study task generalization with meta-RL across manipulation tasks
- simulate two-handed object handover and throw tasks
- compare robot skill learning against human motor development

## When to choose
- you need GPU-accelerated parallel simulation of dual dexterous hands
- you research multi-agent or multi-task RL for bimanual manipulation
- you want a standardized benchmark with diverse objects and tasks

## When to avoid
- you need real hardware robot control rather than simulation
- you require single-arm or non-dexterous manipulation environments
- you cannot use NVIDIA Isaac Gym or an NVIDIA GPU

## Facets
- artifact type: library
- maturity: active
- function: reinforcement-learning, simulation, robotics, benchmarking
- domain: reinforcement-learning, robotics, simulation, machine-learning
- platform: python
- tags: isaac-gym, dexterous-manipulation, bimanual-robotics, marl, gym-environments, sim2real, gpu, linux

## Member repositories
- PKU-MARL/DexterousHands (main) score 35

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:31.452058+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:49:41.533689+00:00, confidence not recorded.
  - readme: https://github.com/PKU-MARL/DexterousHands (fetched 2026-08-28T04:03:31.452058+00:00, sha fa690380d0b1)
  - homepage: https://pku-marl.github.io/DexterousHands/ (fetched 2026-08-29T12:52:44.768594+00:00, sha 283eae1e9366)
- Data as of 2026-08-30T08:39:29.467469+00:00.